That is why Hieronymus AG has created LexMachina, the first NMT engine specializing in the translation of legal and financial texts. Its target users include law firms, banks, (re)insurance companies, consultants and the big4, in Switzerland and Germany. Such companies will benefit from higher quality custom-made NMT engines, with a shorter set-up time.
The purpose of the Datalambic Project, carried out in partnership with iCoSys, is to create a tool ecosystem for semi-automated collection, preparation and correction of high-quality data in order to (re)train neural translation engines in the desired specialization(s), including looped feedback from linguists, lawyers and users. The ecosystem will be modular, including: context-targeted web scraping, document classification, multilingual sentence alignment, term extraction, pseudonymization and customized machine translation post-processing.